The present invention has to do with a method of measuring induced vibrations, over a sequence of periods, in a damped vibratory system and more particularly to the correction of an erroneous DC signal produced by an amplifier, together with low frequency interference, in such a system. One important application of this technology is the measurement of damping.
The measurement of damping, in a damped vibratory or resonant system, finds many application in the industrial arts, among the most important being the measurement of viscosity. One method of measuring viscosity, disclosed in U.S. Pat. No. 5,837,885, involves the perturbation of a fluid by a transducer vibrating near its resonant frequency. The excitation of the transducer is periodically stopped, and after a pause the transducer vibrations are measured. The phase of the received signal is logically “anded” with a phase-shifted signal, which is produced by adding a phase shift δΦ to the excitation signal, producing a control signal which is zero when the two signals are 90° separate in phase. The control signal is used to adjust the excitation frequency in the next iteration. After a number of iterations, typically on the order of 500, with each iteration taking about 1 millisecond, the phase of the received signal should be 90° different from the phase-shifted signal and the frequency should not change from one iteration to the next. The frequency at this state (phase-lock state) is measured. In order that a measurement of damping, and therefore viscosity, can be derived from this frequency, the phase is shifted in sequential periods, each of about one second, by δΦ; −δΦ, and then δΦ again, with the three resulting frequency values being used to determine damping, as explained further in the Detailed Description. The value of δΦ is typically 22.5° or 45°.
The patent discussed above represents a significant advancement in the art of viscosity measurement, with the method and apparatus disclosed gaining widespread acceptance in the field. Nevertheless, possible implementations of this method are limited to a set of applications, which it would be desirable to broaden, although already quite broad.
One set of problems is caused by a DC offset introduced by the amplifier used to amplify the sensed transducer signal. Every amplifier introduces some DC offset, however slight, into its output signal. This DC offset varies with time and temperature. For many electrical devices there is a tradeoff between accuracy and an added cost for expensive components that introduce less DC offset into the system. It is desirable to have a design that permits the use of less expensive components and yet returns a highly accurate result. The DC offset is a potential source of amplifier saturation with attending system nonlinearity, whereas minimizing the potential error caused by the DC offset adds to system complexity.
Unfortunately, the switching between excitation periods and sensing periods in the method of the '885 patent makes it counterproductive to introduce a simple high pass filter into the system to filter out the erroneous DC voltage introduced by the amplifier. Transients that result from the switching process typically have frequency components of frequencies comparable to that of the vibration mode being measured, and thus are passed or even amplified by a conventional high-pass filter. Further complicating the task of reducing the DC offset is the fact that it is the DC offset specifically during the sense periods that should be corrected. Any method not timed to avoid being affected by the excitation period receive signal would risk corrupting the sense period measurement. Not only would such a method not effectively address the problem, but it could even make it worse.
One problematic condition is the measurement of fluid viscosity in an environment that includes a low frequency vibration. This condition occurs in many environments in which it is desired to measure viscosity, for example an industrial or treatment plant in which fluid is being pumped. The pump typically will introduce low frequency vibrations, which may interfere with the frequency measurement, leading to a less certain reading. There is even a possibility that such low frequency vibrations could defeat the phase-lock process, making it impossible to obtain a reading. Low frequency vibrations caused by a physical shock to the viscometer can have the same effect.
Accordingly, it would be desirable to have a method in which an erroneous DC signal produced by a system amplifier and low frequency vibrations caused by ambient noise or a sudden shock to the system could be reduced in amplitude. The frequent switching between excitation and sensing modes greatly complicates the task of originating such a system.